Schizophrenia affects ~1% of all people. It is characterized by delusions and hallucinations resulting in high morbidity and mortality. These symptoms may be underpinned by misperceptions and misinterpretations of sensations that result from a basic inability to predict sensations and adapt to deviations from the expected. If predictive coding mechanisms are dysfunctional, sensations that should have been predicted, but were not, may take on inappropriate salience and lead to the construction of delusional schema to explain aberrant experience. These errors of prediction are costly to society and the patient. Predictive coding models posit that higher order brain areas develop and maintain representations of predictable sensory stimuli, increasing efficiency of neural activity by reducing sensory cortical responses to these stimuli and increasing sensory responses to stimuli that violate the predicted pattern. In such models, the role of primary sensory cortices is to encode and transmit sensory prediction errors, that is, to respond mainly to those stimuli that deviate from the brain's representations of expected stimuli. In context-based predictive coding, the brain acquires and remembers the context of regular temporal and spatial patterns of sensory input. In action-based predictive coding, the predictability of sensory events derives from the fact that sensory events are a predictable consequence of one's own actions, and do not need to be learned or remembered. Conceptually, action-based predictive coding models subsume efference-copy/corollary discharge forward model systems described across the animal kingdom in which an efference copy of an impending motor plan is transmitted from motor to sensory cortex where it generates a corollary discharge representation of the expected sensory consequences of the imminent motor act. Building on our prior work, we propose to ask how predictability and performance monitoring affect action-based predictive coding and how they affect sensory/perceptual processing in schizophrenia. We will use EEG and fMRI, acquired separately and simultaneously, to interrogate details of efference copy and corollary discharge components of action-based predictive coding. Using EEG, we can assess the split-second activity preceding a motor act, or the efference copy. Using fMRI, we can assess the spatial precision of the suppressive action of the corollary discharge mechanism. By integrating EEG and fMRI, we can understand the functional neuroanatomical basis of predictive coding and its abnormalities in schizophrenia. Finally, we are adopting a paradigm and using methods that are translatable to animal models of SZ and that can ultimately be used to identify specific elements of circuits involved in predictive coding, transmitters involved in te neural mechanisms responsible for predictive coding, and even genes that underlie psychotic experiences and behavior.